Operator friction, runtime behavior, and implementation receipts come before the claim.
Research for automation you can defend.
CREATE SOMETHING .io turns experiments, papers, and field notes into a usable research layer for operators. The goal is evidence you can carry into the next build, review, or production decision.
Patterns, benchmarks, and operator notes tied back to real builds.
Papers and field notes stay tied to the workflow, experiment, or policy they support.
Strong patterns can graduate into .space validation or .agency delivery.
published experiments + papers
research categories
featured artifacts to inspect first
database / automation / judgment layers
The research layer should make the next operating decision easier.
This is where CREATE SOMETHING documents what held up in practice, what failed under pressure, and what deserves to be carried forward into the product, policy, or delivery layer.
Patterns start with operator pain, implementation evidence, and runtime behavior before they become a positioning claim.
- Experiments stay tied to the workflow that produced them
- Claims are easier to defend when the artifact trail exists
- Reusable patterns get published only after they survive contact
Measure cost, speed, and maintenance drag across AI-native stacks instead of repeating the same intuition every quarter.
- Cloudflare-native execution and orchestration notes
- Model and framework tradeoffs grounded in implementation work
- Comparisons optimized for operators, not abstract leaderboard chatter
The research output is not just prose. It is policy packs, release checks, contracts, and runbooks that can move into delivery.
- Database / Automation / Judgment is treated as an operating frame
- Evidence rolls forward into specs and policy artifacts
- What gets published should be usable by the next build
This property is tuned for the person who has to explain why a workflow exists, where it breaks, and what should happen next.
- Research is written for implementation and review, not content farming
- Failure modes matter as much as feature lists
- The goal is operational clarity, not thought-leadership theater
Featured Work
Experiments, field notes, and patterns worth inspecting first.
╭──────────────────────────────────────────────────────────╮ │ WEBFLOW ANALYZER LINEAGE │ │ │ │ Jan Feb Mar Apr │ │ detect → extract → govern → productize │ │ experiment MCP policy + reviewer + │ │ origin server review ops creator help │ │ │ │ The system got stronger as its boundaries got clearer. │ ╰──────────────────────────────────────────────────────────╯Webflow Analyzer Lineage: From Detection to Governed Review
researchresearch╭──────────────────────────────────────────────────────────────╮ │ │ │ USER AGENT FILTERS │ │ │ │ "Show me chairs ┌─────────────┐ ┌──────────────┐ │ │ under $2000" ───▶ │ Workers AI │ ───▶ │ category: │ │ │ │ │ │ seating │ │ │ │ Reasoning │ │ price: <2000 │ │ │ │ Streaming │ │ status: any │ │ │ └─────────────┘ └──────────────┘ │ │ │ │ │ │ ▼ ▼ │ │ ╔═══════════════════════════════════╗ │ │ ║ 5 products match ║ │ │ ║ ┌───┐ ┌───┐ ┌───┐ ┌───┐ ┌───┐ ║ │ │ ║ │ │ │ │ │ │ │ │ │ │ ║ │ │ ║ └───┘ └───┘ └───┘ └───┘ └───┘ ║ │ │ ╚═══════════════════════════════════╝ │ │ │ ╰──────────────────────────────────────────────────────────────╯ Ask for what you want. Skip the filter taxonomy.AI-Native Filtering: Natural Language Product Discovery
researchresearchTRADITIONAL SHAPE-AWARE (6D) ╭─────────────╮ ╭─────────────────╮ ╱@@@@@@@@@@@@@@@╲ ╱,,,##########,,,,╲ │@@@@##########@@@│ │,,##MMMMMMMM##,,,│ │@@@############@@│ ──▶ │,#MMMMMMMMMMMM#,,│ │@@@@@@@@@@@@@@@@@@│ │,#MMMMM""""MMM#,,│ ╲@@@@@@@@@@@@@@@╱ ╲,,##########,,,╱ ╰─────────────╯ ╰─────────────────╯ Blurry edges Sharp contours 6D Shape Vector ╭───┬───╮ │ ○ │ ○ │ ← upper (staggered) ├───┼───┤ │ ○ │ ○ │ ← middle ├───┼───┤ │ ○ │ ○ │ ← lower (staggered) ╰───┴───╯ ASCII characters are not pixels— they have SHAPE.Shape-Aware ASCII Renderer: 6D Character Matching
researchresearchN ▲ ▲▲▲▲▲▲▲ ▲▲▲▲▲▲▲▲▲▲▲ ╱─────────────╲ ╱ ○ ○ ○ ○ ○ ○ ╲ │ ○ ○ ● ● ○ ○ ○ │ W ◀│ ○ ○ ● ● ● ○ ○ ○ │▶ E │ ○ ○ ● ○ ○ ○ ○ │ ╲ ○ ○ ○ ○ ○ ○ ╱ ╲─────────────╱ ▼▼▼▼▼▼▼▼▼▼▼ ▼▼▼▼▼▼▼ S ▼ ○ calm ● crowded ● panicked 8,000+ agents │ 60 FPS │ Social Force Model Watch crowds form, bottlenecks emerge, panic spread.Living Arena GPU: WebGPU Crowd Simulation
researchresearch✓ VALIDATED │ v2.3.0 │ 41/41 tests ╔═══════════════════════════════════════════════════════╗ ║ WEBFLOW PLAGIARISM DETECTION ║ ║ ║ ║ ┌──────────┐ ┌──────────┐ ┌──────────┐ ║ ║ │ MinHash │──▶│ LSH │──▶│ PageRank │ ║ ║ │ (1997) │ │ (1998) │ │ (1996) │ ║ ║ └──────────┘ └──────────┘ └──────────┘ ║ ║ │ │ │ ║ ║ └──────────────┴──────────────┘ ║ ║ │ ║ ║ ╔══════▼══════╗ ║ ║ ║ Bayesian ║ ║ ║ ║ Confidence ║ ║ ║ ╚══════╤══════╝ ║ ║ │ ║ ║ ╔══════▼══════╗ ║ ║ ║ MCP Tools ║───▶ Team AI Agents ║ ║ ║ (10 tools) ║ ║ ║ ╚═════════════╝ ║ ║ ║ ║ 9,593 templates │ 517,850 functions │ $2.20/month ║ ╚═══════════════════════════════════════════════════════╝ Classic algorithms. Agent-native delivery.Webflow Plagiarism Detection: Agent-Native Algorithms
researchresearch╱╲ ╱╲ ╱╲ ╱ ╲ ╱ ╲ ╱ ╲ ╱ 🛡️ ╲ ╱ 💡 ╲ ╱ 🌡️ ╲ ╱SECUR╲╱ LITE ╲╱ HVAC ╲ ╱────────────────────────╲ ╱ ╲ │ ╭────────────────╮ │ │ │ │ │ │ │ ╭──────╮ │ │ │ │ │ AI │ │ │ │ │ ╰──────╯ │ │ │ │ │ │ │ │ │ ▼ │ │ │ │ 👤 │ │ │ │ HUMAN DECIDES │ │ │ ╰────────────────╯ │ ╲ ╱ ╲────────────────────────╱ ╲ 📅 ╱╲ 📡 ╱╲ ╱ ╲ ╱ ╲ ╱ ╲ ╱ ╲╱ ╲╱ ╲╱ The system helps. A human chooses.Living Arena: AI-Native Automation at Scale
researchresearch
.io does the reading so the rest of CREATE SOMETHING can move faster.
Research only matters if it transfers cleanly into practice, delivery, or philosophy. That handoff is the point of the network.
Use the workbench to try the idea, inspect the runtime, and see whether the pattern survives execution.
Operationalize the patternMove from research into governed workflow delivery when the operating path becomes commercially or reputationally important.
Contextualize the thesisSee the editorial and philosophical layer that frames why creation matters more than commodity consumption.
Let the research tell the visitor where to go next.
Use .io to understand the evidence, move to .space when the pattern needs runtime validation, and move to .agency when the workflow is ready to be scoped.
- 01 .ltd Canon Clarify the principles, standards, and judgment that should guide the work.
- 02 .io Research Read the evidence, patterns, and operating notes that make the claim defensible.
- 03 .space Workbench Try the routes, tools, and runtime behavior before the pattern becomes delivery.
- 04 .agency Build Turn the fit into a scoped workflow with controls, owners, and handoff notes.
Start with the methodology, then inspect the work.
If you want the operating frame behind the papers, start with the methodology and then move into the experiment and paper archive.
Start with the methodology behind the papers before treating an insight as reusable.
Use the research graph to see how artifacts, claims, and implementation notes connect.
Move strong patterns into runtime practice or scoped delivery only when the evidence holds.